Unlocking the Power of Data: Part 2 - An Overview of Data Strategy
PALO IT / Quinlan & Associates
7 mins read
Data strategy projects begin, first and foremost, with a solid foundation of business planning and an understanding of the company’s long-term goals. This means delving deep into the rationale behind a company looking to utilise data for business-critical decisions and exploring why it may be necessary.
A data strategy project that begins with the technology first is doomed to fail as companies will attempt to shoehorn it into their current business practices.
There are several tools that companies can use to determine the size and scope of a project while bringing lasting adoption of data to an organisation. These are: the project classification, the BAIT (Business, Application, Information, Technology) framework, and change management strategies (see Figure 8).
Many executives may not have a clear picture on how much work goes into a data strategy project. To address this, we have developed a classification system that matches the size and scope of a project, along with the capital and resources needed, to determine how large of a commitment each type of project requires. The biggest first cut would be how mature or clear the current business strategy is, given that technology is only an enabler to a company’s broader strategic objectives.
There are several key considerations that could affect the type of data project selected: impact to business, available resources, time required, and existing infrastructure.
Impact to business stipulates to what degree a given issue affects the organisational strategy as a whole. For example, a simple optimisation of a specific data system would not affect an insurance company’s overall direction. However, the shift to a digital-first company would drastically change the firm’s business strategy and indicate the need for a large-scale strategic project.
Resources, whether manpower or financial, are a key limitation for many management teams, as full transformations require considerably more capital than optimisation projects. Time is another factor to consider; if a rapid deployment or patch to an existing mission critical system is necessary, then optimisation of existing systems might be more desirable rather than longer transformation projects.
Finally, existing infrastructure should be taken into consideration as some data systems may be beyond salvaging, making it more costly to optimise than to transform.
As the name suggests, strategic projects steer the entire business. These are more commonly known as ‘digital transformation’ projects, given their sheer size and scope of work. To be effective, these projects must start with defining and solidifying a company’s current position and strategic goals, while also considering factors that are integral to its long-term future. Subsequently, firms need to develop a list of objectives and supporting goals, which come with a budget. This informs the organisation’s supporting technology and data system choices. We cannot stress enough that this should never be done the other way around.
As an example, think of a traditional insurance company. Many incumbent players still operate under paper claims or laborious manual digital claims, which typically take 2-3 weeks before
customers hear back from the insurance company or receive their pay-out. This represents an unacceptable user experience in today’s digital world and insurance companies know this; their manual processes and suboptimal internal systems are limited because they were not built from the outset to quickly compute and manage claims (or manage follow ups automatically). These manual processes create significant overheads and minimise the company’s ability to create new, tailored insurance products, limiting their revenue growth potential.
Insurtech companies with none of the legacy systems are free to run circles around the traditional insurance industry players, offering microinsurance or tailored premiums based on granular risk calculations, allowing them to tap into a long tail of customers and process claims in a rapid fashion. As a matter of long-term survivability, it is becoming clear that traditional insurance companies are being forced to invest in their digital capabilities before digitally native InsurTechs eat their lunch for good.
A traditional insurance company looking to make a leap into the digital space quickly to defend their market share is a large strategic goal; one that requires complete organisational buy-in, significant investments, and an overhaul of existing systems. Hence, business analysts and IT specialists alike would be involved in the complete BAIT process, set tactical and operational project scopes, and couple it with in-depth change management strategies (along with robust risk management strategies) to ensure implementation success.
Tactical projects are a step below strategic, encompassing lower-level organisational goals that bring a company closer to its strategic vision. They also include the structuring of operational projects and goals.
While it may be tempting for companies to deploy tactical projects, due to their limited scope and relatively smaller budgets, we recommend against this if the company’s broader business strategy has not been established. This is one of the core reasons why dark data and wastage around data investments occur. All too often, we have seen executives charge headfirst into a tactical project to solve a smaller business goal, which ends up conflicting with other legacy systems or current practices without clear demarcations of why a system is being implemented at the highest level of the organisation.
In keeping with the insurance company example: a tactical project would be akin to the company looking to develop a comprehensive risk calculation engine based on collected historical data and other third-party sources to create tailored premiums and bespoke products for their long-tail customers. In this case, the strategy is already well thought out, with its budgetary and regulatory limitations set, so the project would only include the AIT portions of BAIT, simplifying the project and reducing the resources needed to implement it. Some activities would include build or buy considerations, along with integration with the company’s data catalogue and other pre- existing systems.
Operational projects are largely focused on process optimisation. They typically occur when a company’s data fundamentals are already squared away and most of the relevant
infrastructure is in place. Relative to strategic and tactical projects, operational projects are considered the smallest in terms of scope and are more technically focused.
Looking again at the example of an insurance company wanting to create a risk calculation engine, this would operationally require a data filtering system that sieves the relevant information needed for the risk engine to create a risk score. If the company already has a data filtering system in place but experiences latency issues when parsing information to the final risk engine, the project’s scope would be limited to determining the root cause of the latency issue, optimising for costs associated with the delay and identifying the optimal technical solution(s) required to address it.
BAIT is a framework that helps to showcase how a data strategy project should be approached from formulation to execution, while detailing the individual components necessary for a successful project (see Figure 9).
The BAIT framework consists of four parts, all of which represent a vital step in the overall process:
The BAIT framework begins with business problem diagnosis. This is one of the most important steps in the process, as it requires understanding the fundamental nature of the problem the business faces. For example, is the organisation looking to optimise a current practice or refine a feedback loop? Or is it looking beyond its current capabilities to create new business lines or business models?
Organisational resources and obstacles need to be identified in this step. This requires Business Analysts / IT Specialists to work alongside senior management in formulating objectives based on the company’s ambitions and identify obstacles early to prepare for the subsequent steps. Failure to meet analyses of objectives, legal issues, or costs that apply to an organisation could cause serious delays to the project and/or deliver subpar results for the given investment.
The Application stage involves translating business requirements into technical functionalities, to be fulfilled by applications developed by the IT consulting team. The application stage serves as an interface between the business and its information, meant for viewing and performing operations on data as per business requirements. Additionally, this is the stage where proof-of- concepts are built and tested with the company, creating a feedback loop for refinement, ensuring the outcome is tailored to stakeholder and business needs.
The Information stage takes the technical capabilities needed from the Application stage down to the data modeling level. This involves the complete discovery and mapping of a business’s current data locations (including rediscovering previously mentioned dark data hidden throughout the organisation), current data quality, its owners, and the current interoperability of each business unit’s data systems and operations. This is done to create the relevant data models, systems, maps, and data platform designs that are needed to inform subsequent technology selection. If executed holistically, an organisation would create data governance policies and systems that produce trusted data (i.e. information which is immediately usable for analysis by end business users) for anyone in the company.
Last, but certainly not least, the Technology stage involves matching the most suitable technologies to the applications and data platform designs, as stipulated by the Application and Information stages.
It is important to note that this would be within the Business stage’s scope of timeline, budgeting, existing infrastructure, and objectives, which need to be fulfilled. In essence, technology needs to be considered last and depends on all of the previous stages. Once again, it is important to stress that technology is merely a tool to help businesses meet their objectives; executives should be wary of “quick fixes” or “promises” that frontier technologies can bring.
Change management strategy
Tying together a data strategy requires more than just the most advanced hardware, software, or talent; it requires a unifying change management strategy.
Transformations and changes often face resistance from anyone subjected to it, and a data strategy project is no different. Without it, a data investment may not reach its full potential as employees retain old habits, failing to utilise new data tools and processes made available to them.
As such, it is crucial for an organisation to develop appropriate policies and incentives and create a data-centric culture, reinforced with clear governance models. This would encompass four key areas: (1) a communications strategy from the top; (2) actionable frameworks for employees; (3) incentive schemes to promote best-practice data procedures; and (4) agile governance to reflect a fast-changing regulatory environment.
Every change requires a champion within an organisation to see it through. Senior management should endorse and sponsor an entity to oversee the change management strategy with autonomous decision-making powers. This is typically known as an internal enablement committee. Enablement committees are needed at every stage of the project; from sponsoring changes to signing off on budgets to monitoring the overall progress of the company’s data strategy. Both the committee and change management strategies are further elaborated in Section 9: “Change Management Strategy”.
Companies need to carefully tailor their data strategy with these three tools in mind to match their business objectives. This should be paired with near perfect execution of holistic data systems for all business decisions, as doing otherwise would result in, at best, a loss in productivity or, at worst, direct monetary losses.
While large-scale data projects may not be a frequent occurrence for many companies, it is important for firms to engage specialists, who can help steer companies away from common pitfalls when implementing data projects at scale, ensuring the best chance of successful delivery.